[BUG] Multi vector does not return correct result

[heemin32]

What is the bug?
Multi vector with faiss engine does not return correct result in OpenSearch 2.13 and 2.14.

How can one reproduce the bug?

Create index using faiss

PUT /my-knn-index-1
{
  "settings": {
    "index": {
      "knn": true,
      "knn.algo_param.ef_search": 100
    }
  },
  "mappings": {
    "properties": {
      "nested_field": {
        "type": "nested",
        "properties": {
          "my_vector": {
            "type": "knn_vector",
            "dimension": 3,
            "method": {
              "name": "hnsw",
              "space_type": "l2",
              "engine": "faiss",
              "parameters": {
                "ef_construction": 100,
                "m": 16
              }
            }
          }
        }
      }
    }
  }
}

Ingest document

{ "index": { "_index": "my-knn-index-1", "_id": "1" } }
{"nested_field":[{"my_vector":[1, 1, 1]}]}
{ "index": { "_index": "my-knn-index-1", "_id": "2" } }
{"nested_field":[{"my_vector":[2,2,2]}]}
{ "index": { "_index": "my-knn-index-1", "_id": "3" } }
{"nested_field":[{"my_vector":[3,3,3]}]}
{ "index": { "_index": "my-knn-index-1", "_id": "4" } }
{"nested_field":[{"my_vector":[4,4,4]}]}

Query

{
  "query": {
    "nested": {
      "path": "nested_field",
      "query": {
        "knn": {
          "nested_field.my_vector": {
            "vector": [
              4,
              4,
              4
            ],
            "k": 1
          }
        }
      }
    }
  }
}

Response

It should return document 4 with vector [4, 4, 4] but it returned document 1 with vector [1, 1, 1]

{
  "took": 3,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 1,
      "relation": "eq"
    },
    "max_score": 0.035714287,
    "hits": [
      {
        "_index": "my-knn-index-1",
        "_id": "1",
        "_score": 0.035714287,
        "_source": {
          "nested_field": [
            {
              "my_vector": [
                1,
                1,
                1
              ]
            }
          ]
        }
      }
    ]
  }
}

Query with k 2

{
  "query": {
    "nested": {
      "path": "nested_field",
      "query": {
        "knn": {
          "nested_field.my_vector": {
            "vector": [
              4,
              4,
              4
            ],
            "k": 2
          }
        }
      }
    }
  }
}

Response

When k is 2, it returns one correct vector [4, 4, 4] but still has wrong vector [1, 1, 1]

{
  "took": 9,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 2,
      "relation": "eq"
    },
    "max_score": 1.0,
    "hits": [
      {
        "_index": "my-knn-index-1",
        "_id": "4",
        "_score": 1.0,
        "_source": {
          "nested_field": [
            {
              "my_vector": [
                4,
                4,
                4
              ]
            }
          ]
        }
      },
      {
        "_index": "my-knn-index-1",
        "_id": "1",
        "_score": 0.035714287,
        "_source": {
          "nested_field": [
            {
              "my_vector": [
                1,
                1,
                1
              ]
            }
          ]
        }
      }
    ]
  }
}

When k is 3, it returns two correct vectors, [4, 4, 4], and [3, 3, 3] but still has wrong vector [1, 1, 1]

What is the expected behavior?
The search results contain one invalid value for every query.

What is your host/environment?
OS 2.13